Continuous Software Process Assessment: Measurement and Modeling

This report explores the critical aspects of software process assessment, emphasizing the importance of measurement and modeling for effective software engineering. It delves into the challenges associated with traditional assessment methods, highlighting their cost, disruption, and infrequency. The research introduces an approach that combines goal-oriented dimension and a framework for continuous assessment of the software program engineering process. The study examines the use of assessment tools and techniques within an industrial setting, focusing on organizations with ISO 15504 levels one to three. It also discusses the role of various standards and models, including CMM-SW and ISO 15504, in evaluating software processes. The report incorporates concepts like the Goal-Question-Metric (GQM) approach and the Software Engineering Institute's (SEI) Capability Maturity Models (CMMs) to provide a comprehensive understanding of software process improvement and assessment methodologies. The research presents practical methods and tools for measurement data management and assessment profile monitoring, emphasizing the need for continuous assessment and improvement.

Measure and model of software
engineering
Submitted by:

Date:
Introduction
Software program process assessments are regularly used by the software
program sector to evaluate application procedures before instigating
improvement activities. They are moreover used to measure the capability of
a company to create software. Since assessments are usually perceived as
expensive, disruptive plus time-consuming for the workplace there exists a
need to find alternative procedures for software process evaluation.
Interesting for this research has been to understand and specifically
enhance the way an excellent organization may monitor the software
process position between regular assessments and exactly how this
monitoring can be achieved feasibly
in a good industrial environment.
This approach combines goal-oriented dimension and an emerging regular
for software process evaluation as the background framework intended for
continuous assessment of the software program engineering process.
Software program equipment has been created to support the particular
approach that has been tested within an industrial setting. The total
outcomes show that the proposed method is feasible and useful, and
provides new findings plus possibilities for software program procedure
assessment.
Software has become a growing number of pervasive in our society. The
entire year 2000 bug is a good example of the grip software application is
wearing the worldwide economy. Since the need for improved functionality,
high quality, and reliability continues, much better ways to control computer
software growth are needed. Furthermore, as Drain (1999, g. 45) produces,
”aside from being a main cause of disappointment and tension, software
failures have turn out to be a major source of financial strain on businesses
at a time whenever cost containment is the fundamental concern of senior

managers”. Within the 1990´s the software process local community grew
with the importance associated with software in the industry (Humphrey
1999). The paradigm of the software package process proponents is that the
high quality of software program development procedure is in close
relationship using the quality of the making software program (Humphrey
1989, d. 13). There are good examples and proof that an improved software
development process also increases efficiency and reduces variation of the
program development process (Humphrey tout autant que al. 1991; Clark
1997; Herbsleb et al. 1997). Krasner (1999, p. 151) writes: “In a mature
software program organization, the following holds:
o Quality is defined plus predictable therefore
o Expenses and schedules are anticipated and met normally
o Processes are defined plus under statistical control”.
Range of the extensive research
Based on Curtis et al. (1995, p. 10), there are 3 different success criteria
within software program engineering: process, people and technology. This
particular research deals mainly with all the process aspects of software
program executive. Within process oriented analysis there are four main
methods for computer software process enhancement (Kuvaja et al. 1994, p.
29): assessment (e. g. Humphrey 1989), modeling (e. they would. Kellner
and Hansen 1988), measurement (e. g. Basili ET ing. 1994b) and technology
move that can be used independently or even jointly. Of these four, process
modeling and technologies transfer is certainly not part of this research even
though have a place within the enhancement work. For example, process
modeling is needed to review existing procedures - comprehensive process
modeling - or even to requirement new processes to be used: prescriptive
process modeling (Curtis et al. 1992; Bandinelli et al. 1995). This really is the
integrated use of dimension and assessment, which types the general scope

of this thesis. More specifically, this strenuous research involves practical
methods that will support assessment along with measurement data. This
particular would appear easy to come up with specs for any state-of-the-art
integrated process assistance environment that would also provide
information for assessment purposes. Used, it is not really easy to find
organizations that are using integrated procedure support environments. The
choice in this particular research was to recognize the state-of-the-practice in
commercial software program development, and concentrate on how to find
strategies, techniques plus tools for improving examination that would be
useful for the people doing work in the industry today. This thesis mainly
targets organizations associated with ISO 15504 levels one - 3 in maturation.
However, similar techniques are usually apparently being applied within high
maturity organizations even though their experiences are much less often
made public. More information upon high maturation organizations has been
made available as the recent study by Paulk ET ing.
According to the Software program Engineering Institute (SEI), the real
number of organizations starting SPI continues to increase, and half the
nearly organizations confirming size for a software staff be had by the SEI
database of 100 individuals, or less (SEMA 2000). The SPICE assessments
record organizational units amongst ten and 500 THIS PARTICULAR
employees from different organization areas participating in trialing a
software procedure assessment standard (SPICE 1998). Nevertheless, as the
interest broadens and experiences from SPI grow even, little continues to be
reported from the development of software program assessment practice.
The team-based assessment remains as the typical means of assessment,
although the requirement for alternative practices for software program
process assessment has been recognized (Campbell 1995; Miyazaki tout
autant que al. 1995).
Research issue

This extensive research is inspired by the problems with software program
process investigations, Which are documented in the literature and in
exercise. Although assessments are usually considered useful, there are
elements in some of the current techniques that make assessments
perceived to become too costly, disruptive, occasional or rigorous.
Research strategy
This intensive research might be recognized as applied study in the field of
software program engineering. A lot more specifically the research approach
had been helpful research (Järvinen 99, Gary the gadget Guy. 59) that
includes conceptual development, technical solutions, plus their evaluation.
Glass (1994) calls this approach “the executive method”. Trochim (1997) will
abide by this constructivist approach plus claims this represents the
particular post-positivist thinking in modern science.
Research methods
The investigation field is changing really quickly as the software procedure
community is constructing guidelines and frameworks for software program
processes and their evaluation. The development of the recent
INTERNATIONALE ORGANISATION FÜR STANDARDISIERUNG standards ISO
12207 plus ISO 15504 is a good sort of this. The ISO 12207 (1995) points out
the methods needed to fulfill the requirements from the standards - the best
procedures. However, some of these practices are thought applicable only in
a particular situation from the viewpoint from the emerging standard ISO
15504 (1998). Then, there is the hard work to collect the body of information

in software engineering (Bourque et al. 2000) that will seems not to be
completely consistent with ISO 12207 plus ISO 15504.
Conceptual evaluation
The state-of-the-art of software procedure assessment has been analyzed
making use of both literature and energetic involvement in the field as an
assessor and in the
Development of The recognized troubles has led to the characterization of
challenges for IS USUALLY quality (Paper I) as well as the typology of the
different assessment types and modes. Better understanding of industrial
application dimension (Paper IV) assisted to find out new possibilities and the
preliminary idea of measurement centered constant assessment (MCA)
(Paper III). MCA has been processed in the PROFES project (Papers V, VIII)
and included to the PROFES improvement technique (Papers VI, VII) plus
(PROFES-Consortium 2000).
Constructive duties
A technique for using measurement based continuous evaluation (MCA) is
developed (Papers V, VIII). Tools regarding measurement info management
(VTT 1999), mapping measurement information to reference framework (VTT
2000) and assessment user profile monitoring (Etnoteam 1998) along with
several templates have recently been built to support the suggested method
(PROFES-Consortium 2000). In addition, the MCA technique is incorporated
into the PROFES improvement strategy (Papers VI, VII) plus (PROFES-
Consortium 2000), and the POPULARITY assessment methodology (Beitz and
Järvinen 2000).
Research procedure and limitations
This analysis covers work between in the area of software program process
improvement and evaluation. Early initial ideas for that dissertation came in

1994: 1996 when the author began to be active in Bootstrap plus CMM
assessments and their particular comparison (Järvinen 1994b), getting
involved in the software program process evaluation standardization work as
a primary associate of the assessment device team in the SPICE task and the
product manager for that Bootstrap assessment tools (Järvinen 1994a). Mary
Campbell, the particular assessment instrument team head in the SPICE
project provides summarized many of the needs regarding alternative
practices in (Campbell 1995). Along with that time the eye was more focused
on learning the extent of automatization of assessment and associated tool
support. Later, both these particular areas became associated with less
importance to the concentrate of this thesis, however maybe they would
have had more importance if SPICE had stayed the closed, tightly
concentrated standard as was initially recommended (Dorling & Simms 1992;
SPICE-5 1995).
Techniques Assessment
The plan is defined as "An interdependent group of people, objects, plus
treatments constituted to acquire defined goals or a few operational perform
or carrying out specified features. A complete program includes all the
associated products, facilities, materials, pc applications, firmware,
specialized information, solutions, and staff required for functions and
assistance to the diploma necessary for self-sufficient use in the intended
environment". Thus, the master plan is comprised of more than software
package. In order to intelligently specify consistency requirements, we must
understand the specific context in which the software will function.
Software program Design
Considerably, software designs (and analysis) are being expressed as
product oriented paradigms. In these circumstances, this would be essential
for software package engineers to learn the Solitary Modeling Language

[MUL97] and its various design diagrams (e. g., make full use of case,
situation, sequence, state-chart), not only as the style device but
furthermore as the way of identifying where important dependability
requirements can be found inside the system. For example, explaining
customer cable connections with the objects from the system with use
scenario circumstances and identifying the specific states and state
adjustments with state-charts can assist inside identifying the operational
account (see below).
CMM-SW
The program Engineering Institute’s Capability Maturation Models (CMMs)
have
Already been developed in recent years for numerous purposes and there is
an excellent integration effort to harmonies all CMMs under 1 framework -
the CMMI (2000).
However, CMMI provides received criticism (Pierce 2000) for example for
being too Big and complex (437 procedures, 8 KPAs at Degree 2 and 11 KPAs
at Level 3). The state CMMI was not available upon the right time of this
analysis. Hence, only the existing established software program CMM, the
CMM-SW, was chosen for further assessment.
ISO 15504
The LIVEN project was organized in order to harmonies efforts for computer
software process assessment after a good ISO study report (Dorling & Simms
1992) figured there is the need to help the repeatability and assessment of
assessment results. The end result is ISO 15504 -- a framework for software
package process assessment, which is going through a trial time period
before it really is considered to be currently published being an international
standard. Recent advancements show that ISO 15504 may be integrated

more firmly with other related standards (Nevalainen 2000), which may
affect the type and shape of the final worldwide standard. However, these
studies based on the baseline documents accessible upon the time of
creating.
Software measurement concepts
Dimension is the use of metrics to give the value from the aspect scale to an
attribute or even entity (ISO/IEC-8402 1994). Organizations are the objects of
interest intended for attributes and measurement would be the properties of
the company. The particular measurement scale type specifies the
characteristics (Table 2) associated with suitable measures and evaluation
techniques (Kitchenham 1996).
Software program measurement is the continuous procedure for defining,
collecting, and analyzing data on the software package advancement
process and the items to understand control and optimize the process and its
products (Fenton & Pfleeger 1996, pp. 14 - 15).
Software program measurement in practice - GQM
Software measurement furthermore must be systematically planned and
performed (Grady & Caswell 1987). A measurement program much more
likely to be successful if it
is founded on organizational and project objectives (Basili & Rombach 1988).
Goal-Question-Metric (GQM) is really a goal-oriented approach for creating
and running measurement applications that addresses the unique character
of software also dimensions. It has been chosen for this analysis for this good
reason and because nowadays most extra goal powered approaches for
software dimension in use (Bassman et ing. 1994; Park et 's. 1996; Pulford et
's. 1996) seems to be adapted through GQM. The approach will be routinely

mentioned in publications on software additionally dimension as a model of
recommended approach for software dimension (Kitchenham 1996; van
Solingen & Berghout 1999). System details and discussion upon GQM and its
component included in the PROFES improvement methodology within Paper
VII.
Improvement strategies
It is a common misunderstanding within the contemporary software process
local community that assessment methods make up the improvement
methodologies. However, assessment only provides details of the existing
situation to aid improvement planning, and occasionally to support
enhancement monitoring. With that in mind it must be conceded that
specifically those assessment methods which are based on the notion of
maturation (or capability) (Humphrey 1987) tend to lead their consumer into
thinking that the higher the particular maturity, the better, thus
recommending the path for improvement. Nevertheless , Weinberg (1992, p.
31) claims that "maturity is not really the right word for sub cultural patterns
because it indicates brilliance when non-e can be inferred". Weinberg goes
on saying, "The quest for unjustified perfection is not really mature, but
infantile" (ibid, p. 21).
Discussion
Software program process assessment and enhancement, and software
aspect are most active fields in the software program engineering
community, both in the particular academia and in company. In the past ten
to fifteen yrs whole new disciplines have already been established and much
has been accomplished. On process improvement generally there are very a
few great, reported experiences (Humphrey ain is. 1991; Deutsch 1992;
Willis 1998; Ferguson 1999). On the other hand, there are some well-

recorded shortcomings and limitations from the software process ability
centered approaches (Bollinger & McGowan 1991; Card 1991; Humphrey
1991; Card 1992; Weinberg 1992), and there are people who select heroes
over software program processes (Bach 1995).
Generally, there seems to be a tendency to advance from assessment based
improvement to measurement based improvement. When organizations start
to enhance, general guidance that is supplied through assessment, is
enough. Because they continue to improve, businesses require a deeper
understanding of their own procedures and products. To do this they require
sizing data, which assists them find improvement places. This is also visible
within the software process capability versions where measurement is in
concentrate at levels 4 plus 5. El Emam (1998) even publishes articles that
record analysis on assessment information shows that the five (or six) levels
of capability might be reduced into two -- process establishment or maybe
dimension as the main focus intended for improvement.
Evaluation types
3 main types of assessments are usually defined here: Overview,
Concentrated and Continuous Assessment (Figure 6). The basic idea within
the typology is that from Summary to Focused to Constant Assessment there
is an increasing frequency and depth, and reducing breadth in the
assessment. Each one of the three main assessment sorts has their own
particular make use of. Typically, organizations start with a summary
Assessment to get an understanding of the current situation with its talents
and disadvantages. Focused Checks are employed to probe the particular
selected processes in more fine detail then. Continuous Assessments could
be integrated with existing dimension programs to monitor the chosen
processes during and after procedure improvement. There are variants of
every of the three main evaluation types, which are also talked about.

Summary Assessment
Overview Evaluation quickly assesses most or even all processes at
extended periods, e. g. every single a few other years. At minimal, the
assessment results may show which processes can be found, yet will not
reveal the ability level of those processes.
Complete Assessment (Exception)
There is permission in the typology, viz. Complete Assessment. It is included
in the typology because it is sometimes done in exercise and even
mistakenly advocated rather than normal Overview Assessment.
Concentrated Assessment
A Focused Evaluation is done to support an improvement plan. Usually, it is
preceded by simply an Overview Assessment that provides tips for Focused
Assessments. These tests are then synchronized along with a good overall
improvement program so that a Focused Evaluation, a snapshot of probably
just one process, is shipped at an appropriate time. In case a Focused
Assessment is done as well early there is the danger that the process may
alter before suggested improvements are usually making the
recommendations causing, obsolete rework and implemented.
Fixed Assessment
The variant of Focused evaluation, Fixed Evaluation covers just one as well
as a few processes regarding prescribed depth, I. electronic. the depth is
decided in advance, e. g. simply by the certification body or even consumer
needs. Fixed Assessment is carried out as a good often review to guarantee
conformance.